I haven't chimed in on this publicly because I think people would be disappointed by my response and criticise me for being negative. It's a great idea, but will be a complete failure in the real world I'm afraid. Even if I write the reference client and server software to use the protocol, no mining hardware out there and virtually no pool will be using it, thanks to Bitmain's hardware and pool dominance. I've had this discussion with Matt but he's more hopeful than me; otherwise he wouldn't have gone to the effort of formalising his proposal. I'd be happy to be proven wrong on this...